Group Abstract Group Abstract

Message Boards Message Boards

Symlets wavelets give different results after upgrading Mathematica 11.0.1?

Posted 9 years ago
Attachments:
POSTED BY: Claude Mante
5 Replies

If I could debug this for you, I would. But, I cannot reproduce the problem. The problem is somehow specific to your computer.

Please:

(1) Create a new, empty notebook.

(2) Put this code in the notebook.

Don = {....}
TransfoBis = StationaryWaveletTransform[Don, SymletWavelet[8]];
Filtre = InverseWaveletTransform[TransfoBis];
ListPlot[Don - Filtre, PlotRange -> All]

(3) Reduce the size of Don as small as possible while showing your problem.

(4) Evaluate the command SystemInformation[] in this notebook. This command produces a dialog with lots of information about your computer and Mathematica installation.

(5) Send an email to Wolfram Technical Support (support@wolfram.com or https://www.wolfram.com/support/contact/email/?topic=Technical). Attach this new notebook to the email. In the email, tell them that InverseWavletTransform doesn't appear to be working correctly on your computer. They will be able to forward your email to someone who will be able to help you.

POSTED BY: Sean Clarke

Hello, I extracted the fist 100 values of the original list, and the main problem is met again: the wavelet transfom with symlets is not invertible!

More precisely, it depends on the number p of vanishing moments of SymletWavelet[p]: it seems that the transfomation is invertible only for p<4. Is there a reason for that?

I'm working of Window7 professional.

Regards, Claude

Attachments:
POSTED BY: Claude Mante

To begin with, try to make the example as small as possible. Your problem is demonstrated in the first notebook by this:

Don = {....}
TransfoBis = StationaryWaveletTransform[Don, SymletWavelet[8]];
Filtre = InverseWaveletTransform[TransfoBis];
ListPlot[Don - Filtre, PlotRange -> All]

How small can you make "Don" and still reproduce the problem?

Also, I'm not seeing this issue on my machine. For me, The mean value of the absolute value of the residuals is rather small:

Mean[Abs[Don - Filtre]]

5.06531*10^-14

(1) I would suggest trying to reset Mathematica to its installation default configuration according to this article: http://support.wolfram.com/kb/12464 (2) What OS are you working on?

POSTED BY: Sean Clarke

Unfortunately, the previous version was uninstalled from my PC, but I confirm that I noticed that the obtained results were different after upgrading. In fact, I used Symlets for filtering signals, and, as you can see on the second notebook, the signal filtred with these wavelets is completely different from the signal filtered with Haar wavelets, or Daubechies'. I don't think it's normal; in addition, in the notebook I sent yesterday, we could observe that, in this system, InverseWaveletTransform[StationaryWaveletTransform[data, SymletWavelet[8]]] is totally different from data! It's abnormal, I think...

Attachments:
POSTED BY: Claude Mante
POSTED BY: Sean Clarke
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard